New technologies combining digital media item players with dedicated software, together with new media distribution channels through networks are quickly changing the way people organize and play media items. As a direct consequence of such evolution in the media industry, users are faced with a huge volume of available choices that clearly overwhelm them when choosing what item to play in a certain moment.
This overwhelming effect can be easily detected in the music arena, where people are faced with the problem of selecting music from very large collections of songs. However, in the future, we might detect similar effects in other domains like music videos, movies, news, etc.
In general, our invention is applicable to any kind of media item that can be grouped by users forming mediasets. For example, in the music domain, these mediasets could be playlists. Users put music together in playlists to overcome the problem of being overwhelmed when choosing a song from a large collection, or just to enjoy a set of songs in particular situations. For example, one might be interested in having a playlist for running, another for cooking, etc.
This invention addresses the problem of helping users navigate through a media item catalog based on a small set of selected media items. This set of selected media items can be seen as an initial set to build a starting point for the navigation experience.
Different approaches can be considered when building systems to help users navigate a media item catalog. The most commonly used is the keyword based search where the user specifies a set of keywords and the system retrieves the set of media items which contain the keywords in their descriptors. Another approach is to consider a search based on metadata. For example in the music arena, a user might be asking to retrieve rock songs from the 90s.
However, many times users do not know what they are looking for. They want to explore the catalog and find interesting items. This observation is especially relevant for media item catalogs with a clear entertainment focus.
Various approaches can be adopted to personalized recommendations. One kind of approach is about using human expertise to classify the media items and then use these classifications to infer recommendations to users based on an input mediaset. For instance, if in the input mediaset the item x appears and x belongs to the same classification as y, then a system could recommend item y based on the fact that both items are classified in a similar cluster. However, this approach requires an incredibly huge amount of human work and expertise. Another approach is to analyze the data of the items (audio signal for songs, video signal for video, etc) and then try to match users preferences with the extracted analysis. This class of approaches is yet to be shown effective from a technical point of view.
Hosken (U.S. Pat. No. 6,438,579) describes a system and method for recommending media items responsive to query media items based on the explicit and implicit user characterizations of the content of the media items. Dunning, et. al. (U.S. patent application Pubs 2002/0082901 and 2003/0229537) disclose a system and method for discovering relationships between media items based on explicit and implicit user associations between said items. The need remains for improved methods and systems to assist users in navigating through media item catalogs with the ultimate goal of helping them build mediasets and/or discover media items that they will enjoy.